concerns regarding radiation protection standards in a post- fukushima era 31 october 2011 1

Concerns Regarding Radiation Protection Standards in a Post-Fukushima Era

31 October 2011

1

Starting Point:Longstanding Fundamental

EPA Principles

1.Cancer Risks Should Not ExceedAcceptable Risk Range of 10-6 to 10-4

2. Drinking Water Should Not Exceed Safe Drinking Water Levels (MCLs)

2

THE OVERARCHING CONCERN

During the prior administration, ORIA & OEM proposed markedly weakening radiation standards-

to levels far outside the risk range and far above the MCLs, placing longstanding EPA fundamental

policies at risk.3

4

These efforts at weakening standards do not appear to have ceased.

Background

5

6

Fundamental Concern

The science keeps finding radiation to be more dangerous than previously assumed, while politics keeps pushing to relax rather than strengthen radiation protection standards.

7

National Academy of Sciences’ Biological Effects of Ionizing Radiation Studies

BEIR V found radiation ~3-4 times more dangerous per unit dose than previously assumed

BEIR VII found cancer incidence risks 35% higher than BEIR V

Yet the radiation standards of EPA and other agencies have generally not been tightened accordingly. Indeed, as in the case of the PAGs, there has been pressure to dramatically weaken standards further.

8

Whereas EPA generally relies on risk for setting standards, NRC, DOE and ORIA use dose, which makes it hard for decisionmakers to readily judge the appropriateness of proposed radiation standards.

Therefore, it is helpful to keep in mind EPA’s official risk estimates for radiation.

Key Cancer Incidence Risk Conversion Factor for Radiation

Current EPA Factor: 1.16 x 10-3/RemSource: EPA “Blue Book,” EPA Radiogenic Cancer Risk Models and Projections for the U.S. Population (April 2011)

based on exposure spread over a lifetime, or exposure to someone of average age.

Now matches closely to findings of the National Academy of Sciences (BEIR VII) [1.141 x 10-3/Rem]9

10

Thus, any cumulative dose of 0.1 rem -- 100 millirem (not 100 mrem/yr, but total accumulated dose) – or greater to someone of average age would be above the upper edge of EPA’s permissible risk range of 10-4.

87 mrem = 1x10-4 risk per EPA Blue Book. [0.087 rem x 1.16 x 10-3 cancers per rem = 1 x 10-4.]

11

Based on EPA Blue Book data for risk associated with age at exposure, however, the risk is greater for exposures at younger ages.

12

EPA Blue Book: Lifetime Attributable Risk by Age at Exposure

13

According to the EPA Blue Book data, risk before age 30 is on average ~2000 cancers per 10,000 person-Gy, or ~2 x 10-3/rem, ~1.8 times as high as for exposures spread over a lifetime (1160 cancers per 10,000 person-Gy, or 1.16 x 10-3/rem).

16

Estimating risk for exposure before age 15 and before age 30. Risks at ages 0, 5, 10, 15, 20, and 30 (2nd column) taken from EPA Blue Book Table 3-12c (prior slide). Risks between those ages are the averages of those intervals. Risk for exposure up to age 30 is ~2052 cancers per 10,000 person-Gy. Blue Book estimates risk for exposure over lifetime at 1160 cancers/10,000 person-Gy. Exposures up to age 30 thus produce ~1.8 times more cancers.

17

Since EPA’s radiation standards (e.g., CERCLA) are generally based on cancer incidence from 30 years of exposure to age 30, this means that any standard higher than ~1.7 millirem per year would exceed a 1 x 10-4 risk, and anything above about 5 millirem per year would exceed 3 x 10-4, based on EPA’s Blue Book figures.

18

0.0017 rem/yr x 30 yrs x 2 x 10-3 cancers/rem = 1 x 10-4 risk

0.005 rem/yr x 30 yrs x 2 x 10-3 cancers/rem = 3 x 10-4 risk

19

Thus, no radiation standard should exceed a few millirem/year, as it would result in risks outside EPA’s longstanding acceptable risk range, based on EPA’s current radiation risk figures.

20

So, rule of thumb, based on EPA’s most recent official radiation risk #s:

~1.7 mrem/year over one’s first 30 years ≈ 10-4 risk

100mRem/yr for 30 yrs would, according to EPA’s own risk figures, result in cancer incidence about two orders of magnitude higher than the upper end of the acceptable risk range. NRC’s general limits are, in fact, 100 mrem/yr.

21

22

Radiation exposure to a female infant, according to EPA, will result in 4-5 times the cancer risk than the age- and gender-averaged risk. (This doesn’t even take into account that the same amount of radioactivity ingested or inhaled can result in a much higher dose in an infant because of the small body size.)

23Source: EPA Blue Book

24

Source: EPA Blue Book

25

So, exposure to 2 rem in a year (the controversial Japanese proposed standard, partially retreated from, and the existing U.S. PAG value for the intermediate period) would result, according to EPA’s official risk figures, in a radiation-induced cancer risk of 2.3 x 10-3 on an age- and gender-averaged basis, or about one in five hundred, an order of magnitude outside the acceptable risk range.

26

BUT, for a female infant, the risk would be 1 x 10-2--one in a hundred of them would get cancer from that dose, based on EPA’s own official risk estimates. This simply isn’t an acceptable standard.

27

And any standard that permits exposure to the general public of doses of more than 20 millirem total in a single year, even assuming no radiation thereafter is ever received, would permit female infants to receive a risk above the 10-4

upper end of the permissible risk range, even if they never again in their life received any radiation.

EPA has historically opposed any radiation standard above 15

mRem EDE/year. For example, NESHAPs were set at 10

mRem/yr, MCLs at 4 mRem/yr, etc.

Anything higher would exceed the EPA risk range, and EPA has declared would be “non-protective”.28

29

Even some of these older standards would no longer meet the risk range, given the increased cancer risks from radiation determined by EPA in the Blue Book and by NAS in BEIR VII.

30

Summary

According to EPA’s own official most recent risk figures:

Any standard greater than a few millirem/year exceeds EPA’s acceptable risk range.

Any standard for a single year that allows doses greater than ~100 millirem to an adult or ~20 millirem to a child exceeds the risk range for an entire lifetime from just that one year’s dose.

ISSUE 1THE BUSH ADMINISTRATION

PROPOSEDPROTECTIVE ACTION GUIDE (PAG)

31

In the last full day before the Obama Inauguration, outgoing Acting Administrator Marcus Peabody transmitted to the Federal Register ORIA’s rewrite of EPA’s PAGs for dealing with a wide range of radiological releases.

32

33

As drafted by ORIA, the revised PAGs would have astronomically increased permissible exposures to radioactivity.

We had written the outgoing Administration urging that it not engage in such “midnight mischief.”

The Obama Administration, a day or two after taking office, and before the PAG could be published in the Federal Register, pulled it back, pending review by its new team.34

More than 2 ½ years have passed and yet the issue is not resolved.

35

The revised Protective Action Guidance would be applicable to any “event or a series of events, deliberate or accidental, leading to the release or potential release into the environment of radioactive materials in sufficient quantity to warrant consideration of protective actions.”

PAG August 2007 draft, p. 1-1

36

PAGs are to apply to a “wide range of incidents,” including transportation events, releases from a radiopharmaceutical facility, contamination at a scrap metal or recycling facility, incidents at research reactors, and incidents and releases at Department of Energy facilities or civil power reactors.

PAG draft p. ES-2, PAG website http://www.epa.gov/radiation/rert/pags.html

37

“A PAG is defined as ‘the projected dose to reference man, or other defined individual, from a release of radioactive materialat which a specific protective action to reduce or avoid that dose is recommended.’ PAG Draft p. ES-2

38

The Proposed PAG established for long-term cleanup a process called “optimization,” by which there would be no health-based cleanup standard but rather one could choose any cleanup level one wished from various “benchmarks.”

39

Those benchmarks included• 0.1 Rem/year• 1 Rem/year• 10 Rem/yearBelow the benchmark, no cleanup would occur.

40

Over the standard 30 year occupancy period EPA normally assumes, doses to the public under those benchmarks would result in cancer risks, according to EPA’s official risk figures (Blue Book), of approximately:

•~7 x 10-3

•~7 x 10-2 and•~7 x 10-1 respectively

These benchmarks are nearly 2-4 orders of magnitude outside EPA’s acceptable risk range, and as high as a 7 in 10 risk

41

42

This does not include a correction for increased risk for exposures prior to age 30. Correcting for age at exposure would roughly double the above risk estimates.

In 2001, EPA’s position was that any revised general PAGs should be based on Safe Drinking Water MCLs for intermediate phase water consumption and the CERCLA risk range for long-term cleanup standards. Subsequently, ORIA effectively abandoned this position and put forward proposals to extraordinarily relax these standards.

43

Intermediate Phase Water PAGsThe intermediate phase lasts for one to several years after the initial release.

Buried deep in the proposed PAGs was the following table—with no explanation that the water contamination limits in the table are different than longstanding EPA water standards, nor any substantive explanation of how they were derived. (The right-hand two columns, marked DRL, or Derived Response Levels, are the proposed water contaminant limits in the draft PAGs.)44

EPA’s existing emergency response levels – the CERCLA Program’s Removal Action Levels (RALs) – are the MCLs.See, e.g., OSWER, Revised Superfund Removal Action Levels,” 17 September 2008, from Deborah Dietrich, Director of Emergency Management, to Regional Superfund Division Directors and Regional Removal Managers

(Note: CBG 2008 study, “Proposed Relaxation of EPA Drinking Water Standards for Radioactivity,” compared the proposed PAGs to the RALs that were in effect prior to the above directive.)

49

EPAMaximum Concentration

Limits (MCLs)/Response Action Levels

(RALs) for Beta and Photon Emitters in Drinking Water

Source: EPA Directive 9283.1-14,Use of Uranium Drinking Water Standards under 40 CFR 141 and 40 CFR 192 as Remediation Goals for Groundwater at CERLCA Sites, 6 November 2001

50

A Comparison of the Proposed Water PAGs and Longstanding MCLs

53

59

PAGs w/ decay compared to MCLs/RALs

63

EPA’s previous method for determining RALs was to take the greater of the MCL or the concentration associated with a 10-4 risk. Using this method, we calculated RALs and compared them with the proposed PAG values. The PAGs were many orders of magnitude higher.

Forcing the public to drink water contaminated at orders of magnitude above Safe Drinking Water Levels or RALs, for a year or more after an event, is contrary to longstanding EPA practice, which is to provide alternative drinking water supplies or require treatment of contaminated supplies.67

68

Current ConcernsPress reports indicate that EPA is moving forward with optimization for long-term cleanup but without use of the term. It is reported that no firm standard for long term cleanup will be specified, leaving it to be determined at the time and permitting it to be picked from various benchmarks. That is optimization without calling it such. EPA should set guidance now, rather than allowing confusion in the wake of an emergency, and that guidance should be consistent with CERCLA’s 10-6 to 10-4 risk range.

69

EPA conduct during U.S. monitoring of Fukushima radiation appeared to in fact follow the withdrawn PAGs, with EPA attacking its own Safe Drinking Water Act MCLs and instead making inappropriate comparisons to levels orders of magnitude more lax.

70

Two years after we raised our original concerns, no one at EPA has talked to us. It appears clear NRC and DOE are in the loop, pushing for weakening the PAGs, but those of us with an interest in protecting public health and the environment have been frozen out. This needs to change.

71

Issue 2: EPA Radnet Monitoring of Fukushima Radiation in the U.S.

72

PAGs are Guidance for Protective Actions to be Taken When Specified Radiation Levels are Exceeded. For PAGs to be effective, the PAG levels must be protective, and there must be a workable system for determining when those radiation levels are exceeded.

73

The EPA RADnet U.S. monitoring system’s performance during the Fukushima accident was a good, real-world test of preparedness for a radiological release affecting this country.

74

The RADnet system performed poorly. Substantial improvements need to be made to assure that were there a major radiological event in the U.S., protective actions could be taken.

75

The Fukushima Daiichi accident began on Friday, March 11. The event resulted in meltdowns in three reactors and damage to up to four irradiated fuel pools. Very large amounts of radioactivity was being released directly into the environment. The accident continues to this day.

76

U.S. policy is that "EPA is theCoordinating Federal Agency for the U.S. government's response to foreign nuclear accidents."

See http://www.epa.gov/rpdweb00/rert/internationalplans.html#nuclearaccident

77

However, at least in the early period of the accident, EPA was apparently relegated to a secondary position in coordinating the U.S. response to this nuclear accident.

78

On Monday, March 14, a press conference was held at the White House with NRC and DOE. NRC stated that no “harmful” radiation could reach the United States.EPA promptly posted on its website an affirmation of this statement.

79

This was disturbing, and has created substantial trouble ever since. The longstanding position of both NRC and DOE has been that there is no safe level of radiation, all radiation increases risk of cancer, and that there is no threshold below which there is no risk.

All agencies, including EPA, accept the Linear No-Threshold (LNT) Model, and the National Academy of Sciences has recently re-affirmed it.

LNT means there is no threshold of radiation exposure below which the risk is zero; risk increases with dose.

80

81

These misstatements have caused great confusion in the press and public, leading both to believe that the EPA’s position is that only very high doses of radiation are harmful and that below those doses there is no risk.

82

As the accident went on, EPA particularly pointed to its RADnet system of stationary air monitors for its claim that all data indicated levels “thousands of times below any level of concern.”

83

However, the stationary air monitor system had major problems.

Less than half the monitors were fully functional at the time of the accident.

Many were broken, and had been broken for months.

84Source: EPA website, 20 March 2011

85

Source: EPA website, 20 March 2011

86

So, according to EPA’s own compilation, less than half of the monitors were “running.”

87

Indeed, they had apparently been broken for many months, without anyone noticing or taking steps to fix them.

For example, the San Diego Union discovered that San Diego’s RADnet monitor had been broken since November.

88

EPA’s response was that it got the broken ones fixed after the accident. But how can one have an emergency system where many of the monitors had been broken for months and weren’t functioning when the accident occurred?

89

Those that did get fixed weren’t fixed until a couple of weeks into the accident. In most accidents, the event would be long over before the monitors even started operating.

91

EPA officials told the press that the stations marked light blue in the EPA map weren’t really broken, just “undergoing data review” that takes an hour or so before posting the data.

However, that doesn’t seem to be accurate. Many of the light blue stations had broken monitors that stayed broken throughout the accident.

93

And others seems to have been working when the accident began, and broke a few days later and never got fixed.

95

But even had the near-real-time parts of the monitors been working, they couldn’t see much that would be important. They are very insensitive, producing only gross counts that require very high increases to be visible.

96

Their primary way of working is by letting the air filter collect particulates for 3-4 days, then mailing them to Montgomery, Alabama for measurement, which takes additional time. Any measurement would thus occur a week or so after the actual radiation exposure.

97

Yet even so, virtually none of the filters had data reported for particular radionuclides. It turns out EPA’s practice is to ship the filters back to Montgomery, but not actually measure them for particular radionuclides unless the gross beta count is 100 or 200 times normal.

98

This means that, for example, if you had a RADnet device in Washington, and if it was working, radioactivity levels in Washington air would have to rise two orders of magnitude before EPA would even measure to see what radionuclides were in the air.

99

But even had the system been fully functional, it was incapable, as operated, of seeing most radioactivity. The stationary air samplers rely only on air filters, and most radioiodine, a key isotope of concern, would pass right through the filter because it is generally in elemental (vapor) form. The RADnet systems is blind to most radioiodine.

100

Radioiodine is a critical radionuclide in nuclear accidents. It is produced in copious quantities; is very volatile and thus released at a high rate from damaged reactor fuel; and concentrates in the thyroid gland and readily produces thyroid cancer. As we shall see, radioiodine was a dominant finding in other media, e.g. rainwater and milk.

101

EPA does have devices that can measure all forms of radioiodine—deployable monitors. These deployables use both filters and charcoal cartridges; the former pick up particulates, the latter the gaseous forms. They also have more efficient filters, and were to be sending filters and cartridges for measurement more frequently, reducing delays in data.

102

There were large gaps in the RADnet system, even if all devices were working. For example, there were no monitors on the coast between LA and San Francisco.

103 Source: EPA website, 20 March 2011

104

Some plume projections indicated the plume would in fact arrive between LA and San Francisco.

105

Source: Washington Post

106

So, EPA commenced steps to place deployables up and down the West Coast, filling in gaps in the system. The deployables were “stages”—sent out to several locations from which they would then be deployed in the field. Cooperation from state air pollution officers was arranged. And then something strange happened.

108

Despite a FOIA request, we have been unable to get any information as to why EPA HQ decided to stop the plans for deploying the deployable monitors.

In the end, they sat, with a couple of exceptions, in the offices and warehouses to which they had been staged, without getting placed in the field where they could provide useful data.

109

Precipitation

Perhaps more important than air monitoring was precipitation. Radionuclides come back to earth in rain, snow, and sleet.

110

They can thus work their way into drinking water and, even more critically, food, since many radionuclides bioaccumulate, i.e., increase in concentration as they move their way up the food chain. Strontium-90, for example, concentrates in grass, in the cow that eats the grass, in the milk, and then in human bone, where it can cause bone cancer and leukemia. Radioiodine has similar concentrating properties.

111

Yet EPA did not address precipitation until Pennsylvania and Massachussetts state authorities released measurements they had made finding quite elevated levels of radioiodine in precipitation.

Then EPA’s measurements also showed elevated radioiodine all across the country, which far exceeded its own Safe Drinking Water levels (MCLs). But EPA downplayed the significance, and seemed to attack its own MCLs.

112

We have in the slides that follow compiled the precipitation detects for I-131 and compared them with EPA’s Maximum Contaminant Levels for drinking water.

117

EPA appeared to attack its own MCLs, arguing misleadingly that they are based on 70-year exposure. In fact, EPA’s Safe Drinking Water standards bar exposure above the MCL for more than a year. Even averaged over a year, these readings would exceed the MCL.

118

Furthermore, note that the bulk of the cancer risk from ingesting I-131 occurs in the first few years. The risk is heavily “front-loaded” in terms of age.

119

Source: EPA “Blue Book”

120

Lastly, as indicated earlier, EPA’s MCLs are in fact the basis for EPA’s existing emergency levels for providing replacement water supplies or requiring treatment– the RALs of your Superfund program. It is also worth remembering that we are talking not about PAG standards for the immediate emergency, but for the year or two after the emergency is presumed to have ended.

121

EPA kept saying that all measurements were orders of magnitude below “any level of concern.” But, of course, the Safe Drinking Water MCLs are EPA levels of concern, and they were far exceeded, as shown in the next slides.

125

Food

One of the main pathways of concern is food. However, there were no measurements of foodstuffs except for some milk measurements, and those were spotty. Virtually no strontium-90 measurements were made, despite its critical importance in the milk pathway. And radiodine was detected above the drinking water MCL.

127

But EPA continued to assert that measurements were far below any level of concern.” It appeared to do so by comparing readings not with its Safe Drinking MCLs but with levels which are thousands of times higher.

128

In essence, ORIA acted during Fukushima as though the Bush-era proposed PAGs were in fact in effect. Rather than relying on MCLs, it used comparisons to de facto PAGs orders of magnitude higher.

129

The normal milk monitoring procedures are troubling. Only a relatively few milk samples are taken, they are only taken quarterly, they are generally composited (so elevated values from a dairy would be diluted and you wouldn’t be able to locate which dairy was the cause of elevated composited readings), and strontium measurements are taken only every 4th year from sample locations.

130

Normal procedure is to not get around to measuring milk samples for gross gamma radiation (e.g. I-131) for several weeks and for strontium-90 for half a year to a year. Obviously, if one found a problem, any protective action would be impossible, as the milk would long since have been consumed.

131

The Fukushima tragedy provided an important test of the adequacy of EPA’s planning and capability of responding to a radiological release affecting the U.S. The problems identified need to be promptly remedied so that were there an event in the U.S., the system would in fact be capable of leading to protective actions that were truly protective.

132

Conclusions

1. The science keeps showing radiation to be more dangerous than presumed before, but standards keep getting weakened rather than strengthened.

133

2. The Bush-era proposed PAGs have not been fixed. Reports indicate that the essence of the problem with them – optimization for long-term cleanup and suggesting far higher permissible concentrations in drinking water than the Safe Drinking Water levels or emergency RALS—continue to be considered, albeit without use of the explicit language from the Bush draft.

134

3. The EPA RADnet system—essential for protective actions to actually be initiated—largely failed to function appropriately during the Fukushima disaster. This should be a wake-up call to fix the system before it may be needed in a domestic nuclear event.

135

4. More than 2 years have passed since we expressed our original concerns, and yet there has been no engagement with the environmental and public health communities on these issues. Yet DOE and NRC, among others, are clearly interacting with EPA pushing for weakening of protections. EPA should be transparent, and bring us into the loop.

136

Fundamental closing thought: As EPA said many years ago in opposing NRC radiation standards as non-protective:

“To put it bluntly, radiation should not be treated as a privileged pollutant.”

137

138

Contact: Daniel HirschCommittee to Bridge the Gap(831) 336-8003

concerns regarding radiation protection standards in a post- fukushima era 31 october 2011 1

Documents